FGGS-LiDAR: Ultra-Fast, GPU-Accelerated Simulation from General 3DGS Models to LiDAR

2025-09-22

AI Summary

Existing 3D Gaussian Splatting (3DGS) models lack direct compatibility with high-fidelity LiDAR simulation, limiting their utility in multimodal robotics and autonomous driving simulation. This paper introduces the first general-purpose conversion framework that transforms any pre-trained 3DGS model into a LiDAR-perceptible geometric representation—without fine-tuning, supervision, or architectural modification. Our method comprises three key components: (1) an unsupervised geometric extraction pipeline leveraging voxelization and truncated signed distance fields (TSDF); (2) a GPU-accelerated ray-casting echo simulation algorithm achieving >500 FPS; and (3) high-accuracy depth map generation applicable to both indoor and outdoor scenes. Experiments demonstrate substantial improvements in geometric fidelity, reusability, and plug-and-play capability of 3DGS assets for LiDAR-based perception tasks in simulation.

Technology Category

• Computer Vision: 3D Computer Vision
• Intelligent Robots: Multimodal Perception & Sensor Fusion
• Search and Optimization: Sampling/Simulation-based Search

Application Category

• User Modeling, Personalization and Recommendation: User modeling and simulation for interactive and conversational systems
• Search and Retrieval-Augmented AI: Large language models for search
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

Abstract

While 3D Gaussian Splatting (3DGS) has revolutionized photorealistic rendering, its vast ecosystem of assets remains incompatible with high-performance LiDAR simulation, a critical tool for robotics and autonomous driving. We present extbf{FGGS-LiDAR}, a framework that bridges this gap with a truly plug-and-play approach. Our method converts extit{any} pretrained 3DGS model into a high-fidelity, watertight mesh without requiring LiDAR-specific supervision or architectural alterations. This conversion is achieved through a general pipeline of volumetric discretization and Truncated Signed Distance Field (TSDF) extraction. We pair this with a highly optimized, GPU-accelerated ray-casting module that simulates LiDAR returns at over 500 FPS. We validate our approach on indoor and outdoor scenes, demonstrating exceptional geometric fidelity; By enabling the direct reuse of 3DGS assets for geometrically accurate depth sensing, our framework extends their utility beyond visualization and unlocks new capabilities for scalable, multimodal simulation. Our open-source implementation is available at https://github.com/TATP-233/FGGS-LiDAR.

Problem

• Bridges 3DGS assets to LiDAR simulation for robotics applications
• Converts pretrained 3DGS models into watertight meshes without LiDAR supervision
• Enables ultra-fast GPU-accelerated LiDAR simulation at over 500 FPS

Innovation

• Converts 3DGS models to watertight mesh
• Uses volumetric discretization and TSDF extraction
• GPU-accelerated ray-casting for fast LiDAR simulation